Introduction

In clinical trials and in clinical practice, the management of adverse events related to the use of a drug very often involves dose adjustment that consists of temporary therapy suspension or temporary dose reductions, with resumption at full dose upon resolution of the adverse event, or permanent dose reduction until the end of treatment.

As a result, not all patients take the standard or the recommended dose; a percentage of them take a lower dose.

In clinical trials, the correspondence between the recommended dose and the dose delivered to the patient is defined as the relative dose intensity (RDI), expressed as percentage. Some studies have shown that a better RDI corresponds to a better clinical outcome [1, 2]; at the same time, however, it would be useful to have more details about the dose intensities of each arm of the trials [3]. Clinical trials often do not provide data on the number of patients who change their drug dose during chemotherapy. These data should be appropriately explained and described [4]. In target therapies, dose selection is fundamental for therapy optimization in terms of risk/benefit [5]. At present, investigators tend to share and make available data from clinical trials in order to conduct meta- or other statistical analyses that help physicians provide better treatments to patients [6]. For this reason, it is essential to have the most complete results available on all the variables involved in patient management, and, by extension, on treatment efficacy and safety.

The objective of this study was to evaluate how much is dose adjustment prevalent in clinical trials of oral targeted therapy drugs for the treatment of solid tumors, as well as how such adjustment is described and whether dose changes are considered in the results as a factor that may influence a clinical outcome.

Materials and methods

We considered all pivotal clinical trials of oral targeted therapy drugs approved by the European Medicine Agency as of July 31, 2018, for the treatment of solid tumors.

We performed a search on the official EMA site on human medicines, using as Keyword Search the ATC Code L01X (other antineoplastic agents); from the list of drugs results, we subsequently excluded antineoplastic drugs for hematological diseases, as well as refused and withdrawn drugs.

We decided to consider only pivotal clinical trials because they yield data that populate information entered in the Summary of Product Characteristics (SpC) and are taken as reference for dose adjustments in clinical practice. We only considered oral targeted therapy drugs for solid tumors because some studies that evaluate dosage adjustment for conventional chemotherapies already exist, though reviews in the literature that investigate targeted therapies are lacking.

Search for articles: we considered all pivotal articles included in section 5.1 Pharmacodynamic Properties, subsection Clinical Studies, of the SpC.

For all analyzed drugs, we recorded full dosages, dose adjustments with relative reduction percentage, reason for the adjustments, number of patients included in the trial, percentage of patients who reduced their dosage or temporarily discontinued therapy, cause of dose reduction, and presence or absence of reference to a clinical outcome in patients who reduced their dose or discontinued therapy.

Results

From an initial search on the EMA site, we identified 125 drugs, of which 8 are in withdrawal status and 6 in refused status; from the remaining 111, we excluded those prescribed for the treatment for hematological diseases, all non-targeted therapies, and all parenteral therapies.

At the end of our screening, we considered 29 drugs: afatinib, alectinib, axitinib, cabozantinib, ceritinib, cobimetinib, crizotinib, dabrafenib, erlotinib, everolimus, imatinib, lenvatinib, lapatinib, nintedanib, niraparib, olaparib, osimertinib, palbociclib, pazopanib, regorafenib, ribociclib, rupacarib, sonidegib, sorafenib, sunitinib, tivozanib, trametinib, vandetanib, vemurafenib.

Based on the SpC, the 29 drugs call for dosage adjustment for the management of side effects, intended both as dose reduction and temporary suspension of treatment. Only 2 drugs (afatinib and axitinib), according to their SpC, allow for the possibility of reducing or increasing dosage vis-à-vis the recommended dose, if the medicine is well tolerated and if patients have not experienced adverse events.

Table 1 shows the recommended dosage, dose increase, dose reduction with the relative percentage compared to the recommended dosage, and administration suspension with relative number of days contemplated in the SpC.

Table 1 Recommended, increased, and reduced doses

Of the 29 drugs which allow dose reduction, only 6 accommodate a single-dose reduction: crizotinib, everolimus, osimertinib, erlotinib, tivozanib, sonidegib. The remaining treatments provide instead at least 2 reductions. On the other hand, all drugs allow a temporary interruption for the management of side effects.

We identified a range going from a minimum 20% reduction of crizotinib up to a maximum 67% reduction, compared to the recommended dose for cabozantinib, ribociclib, vandetanib, ceritinib, and dabrafenib. The maximum dose reduction occurs between the dose increased to 10 mg twice per day of axitinib and the minimum dose of 2.5 mg twice per day, with a 75% reduction between the former and the latter dose.

We considered 74 pivotal clinical trials, for a total of 17,637 patients evaluated (Table 2). Of the 74 pivotal clinical trials we considered, 56 (76%) provide information on the number of patients who reduced their dose, and 41 (55%) on the number of dose interruptions. Eleven of the 74 trials, namely 14.8%, for a total of 2377 patients (12.7%), do not provide information about the dosage adjustment, reduction and suspension, and amount of drug taken by the sample.

Table 2 % reductions and interruptions in pivotal trials

For the remaining trials, although dosage reductions and therapy suspension are allowed, the number or percentage of patients undergoing adjustment is not specified. In 29 of 74 (39%) trials, the authors referred to the dose taken by the sample, considering as an indicator the RDI expressed as absolute or percentage value, the mean or median dose intensity expressed in milligram, or the percent of patients who took more 80 or 90% of the recommended therapy. Dose adjustment is prevalent in clinical trials. From data presented in the articles, it is evident how dose reductions affect 4475 of 13,856 patients (32%); therapy suspension concerns instead 4730 of 10,814 patients (44%).

The causes of dosage adjustment are rooted in the management of adverse events, and, in particular, grade 3 and grade 4 events of hematological and gastrointestinal nature, fatigue, and increase in the level of transaminases.

Table 3 shows the 29 drugs with one or more indications organized in percent of reduction’s rates, for 10% cutoffs.

Table 3 Rates reduction in 10% cutoffs

In reporting results pertaining to treatment efficacy, pivotal clinical trials do not consider subgroup efficacy analysis on the basis of the total dosage taken.

Discussion

Dosage adjustments are mainly related to the management of side effects, and are prevalent in clinical trials. In most trials, the authors were careful to report the number of patients for whom dosage adjustment was necessary. Less than half of the trials report a general measure of drug intake, usually, but not only, expressed as relative dose intensity; as a result, for most trials, it is impossible to determine what amount of medicine was taken by the sample. Consequently, we are prevented from ascertaining how many patients have taken an adjusted versus a recommended dose. No trial allocates patients to subgroups based on drug intake according to a recommended or an adjusted dose; therefore, the trials on file do not provide any data discussing how the clinical response can vary according to the dose administered and, by extension, on the side effects managed with dose adjustment.

For some pivotal clinical trials, post hoc analyses provide such information: for afatinib [81, 82] and palbociclib [83], no difference between full and reduced dose was found in the clinical outcome, whereas for everolimus, a dose-response model showed that the full dose is more effective than the reduced dose, which, in turn, is more effective than placebo [84]. Although pivotal clinical trials are neither conceived nor designed to evaluate differences in clinical outcomes on the basis of varying drug doses, making such information available would increase the quality of the information, making it more complete and providing a database to produce reviews and meta-analyses on this topic.

As for drug use indices, it would be useful to know with certainty how each patient took the drug and what was his/her clinical outcome. Without sharing such significant data, it would be useful, albeit not exhaustive, to report at minimum, for each arm, the average or total median of the dose taken, the percentage of patients taking more medication than the established cutoff (usually between 80 and 90%) and the RDI.

We believe that sharing such data is fundamental in order to produce dose-response models useful for the optimization of clinical therapy and in clinical practice, as it would provide evidence-based information to support efficacy, or at minimum, dosage adjustments.

Conclusions

It would be advisable for pivotal clinical trials to give more relevance to dosage management, which is a widely used tool for the management of adverse events in clinical practice. This information would help clinicians in managing patients with greater awareness, and in relying on evidence-based data. To date, such information is lacking.